Research Scientist / Research Engineer

About the Role

We are a small, fast-moving AI training data company building the next generation of verifier-grounded training data for frontier AI models. Our core conviction is that model capability breakthroughs require data breakthroughs — and we're tackling that problem head-on using formal proof systems, simulators, executable tests, and oracle databases to produce data that is correct by construction .

We work closely with frontier model teams across LLMs, robotics, and AI for Science, injecting high-quality, information-dense content — covering physics, biological facts, and self-consistent logic — directly into AI training pipelines. We're hiring both Research Scientists and Research Engineers to join our founding team in San Francisco.

What You'll Do

• Design and build automated pipelines for generating verifiable, high-quality AI training data using formal methods, simulators, and oracle databases.

• Research and develop novel approaches to producing training data that encodes fundamental natural-world knowledge (physics, biology, logic) in natural language.

• Collaborate directly with frontier AI lab partners to understand data needs and translate them into scalable data generation solutions.

• Run experiments to evaluate the downstream impact of synthetic and verifier-grounded data on model capability and performance.

• Contribute to the technical direction of the company as an early team member with significant scope and ownership.

What We're Looking For

Required:

• Based in San Francisco, CA, or willing to relocate — this is a fully on-site role.

• Strong background in machine learning, AI research, or related engineering disciplines (0–12 years of experience considered across both tracks).

• Experience with one or more of: large language models, formal verification, robotics simulation, scientific computing, or data pipeline engineering.

• Comfort working in a very early-stage environment with high ambiguity and a small team.

• Rigorous, first-principles thinking about data quality, model training, and evaluation.

Nice to Have:

• Familiarity with formal proof systems (e.g., Lean, Coq, Isabelle) or symbolic reasoning tools.

• Experience in robotics, physics simulation, or AI for Science domains.

• Prior work at or close collaboration with frontier AI labs.

• Background in synthetic data generation or data-centric AI research.

Compensation & Benefits

• Salary: $100,000 – $300,000 USD annually, depending on experience and role track (Scientist vs. Engineer).

• Early-stage equity commensurate with founding-team proximity.

• Visa sponsorship: Not available — candidates must be authorized to work in the United States.

Location

• On-site, full-time in San Francisco, CA, United States.

• Relocation support available for candidates outside the Bay Area.